6.5. Miscellaneous

6.5.1. Libadalang.Lexer

This package provides types and primitives to split text streams into lists of tokens.

type Lexer_Input

Discriminants:

Kind (Langkit_Support.Types.Lexer_Input_Kind) –

Components:

Charset (Ada.Strings.Unbounded.Unbounded_String) – Name of the charset to use in order to decode the input source
Read_BOM (Standard.Boolean) – Whether the lexer should look for an optional Byte Order Mark
Filename (GNATCOLL.VFS.Virtual_File) – Name of the file to read
Bytes (Ada.Strings.Unbounded.Unbounded_String) – Source buffer to read
Text (Unbounded_Text_Type) – Source buffer to read

Source buffer to read

procedure Extract_Tokens (Input : Lexer_Input; With_Trivia : Standard.Boolean; TDH : Langkit_Support.Token_Data_Handlers.Token_Data_Handler; Diagnostics : Langkit_Support.Diagnostics.Diagnostics_Vectors.Vector)

Extract tokens out of the given Input and store them into TDH.

Raise a Name_Error exception if this involves reading a file that can not be open. Raise an Unknown_Charset exception if the requested charset is unknown. Raise an Invalid_Input exception if the source cannot be decoded using the given Charset.

function Is_Keyword (TDH : Langkit_Support.Token_Data_Handlers.Token_Data_Handler; Index : Langkit_Support.Token_Data_Handlers.Token_Or_Trivia_Index; Version : Language_Version) → Standard.Boolean

Given an Ada language version, return whether a token is an Ada keyword.

Due to the way Libadalang works, every token added after Ada 83 is lexed as a regular identifier, and then treated specially by the parser in some circumstances (being akin to the notion of reserved word).

This function returns True for regular lexer keywords, as well as for those identifiers.

6.5.2. Libadalang.Iterators

This package provides an interface to iterate on nodes in parse trees and to look for node patterns.

First, as an alternative to Libadalang.Analysis.Traverse, you can do:

declare
   It   : Traverse_Iterator'Class := Traverse (My_Unit.Root);
   Node : Ada_Node;
begin
   while It.Next (Node) loop
      --  Process Node
   end loop;
end;

Now, if you are exclusively looking for nodes whose text is either foo or bar, you can replace the call to Traverse with the following:

Find (My_Unit.Root, Text_Is ("foo") or Text_Is ("bar"));

The Find-like functions below take as a second argument a predicate, which is an object that can decide if a node should be processed or not. This package provides several built-in predicates (Kind_Is, Text_Is, etc.), then you can either define your own, derivating the Ada_Node_Predicate_Interface type, or compose them using Ada’s boolean operators.

package Ada_Node_Iterators is new Support.Iterators

package Ada_Node_Iterators is new Support.Iterators
     (Element_Type  => Ada_Node,
      Element_Array => Ada_Node_Array);

Instantiated generic package:: Ada_Node_Iterators

type Traverse_Iterator: Iterator that yields nodes from a tree

function Traverse (Root : Ada_Node'Class) → Traverse_Iterator'Class : Return an iterator that yields all nodes under Root (included) in a prefix DFS (depth first search) fashion.

type Ada_Node_Predicate_Interface

Predicate on nodes.

Useful predicates often rely on values from some context, so predicates that are mere accesses to a function are not powerful enough. Having a full interface for this makes it possible to package both the predicate code and some data it needs.

Note that predicates are not thread-safe: make sure you don’t use a predicate from multiple threads, as they can contain caches.

function Evaluate (P : Ada_Node_Predicate_Interface; N : Ada_Node) → Standard.Boolean: Return the value of the predicate for the N node

package Ada_Node_Predicate_References is new GNATCOLL.Refcount.Shared_Pointers

package Ada_Node_Predicate_References is new
      GNATCOLL.Refcount.Shared_Pointers (Ada_Node_Predicate_Interface'Class);

Instantiated generic package:: Ada_Node_Predicate_References

type Ada_Node_Predicate

Ref-counted reference to a predicate

function "not" (Predicate : Ada_Node_Predicate) → Ada_Node_Predicate : Return a predicate that accepts only nodes that are not accepted by Predicate.

function "and" (Left, Right : Ada_Node_Predicate) → Ada_Node_Predicate : Return a predicate that accepts only nodes that are accepted by both Left and Right.

function "or" (Left, Right : Ada_Node_Predicate) → Ada_Node_Predicate : Return a predicate that accepts only nodes that are accepted by Left or Right.

function For_All (Predicates : Ada_Node_Predicate_Array) → Ada_Node_Predicate : Return a predicate that accepts only nodes that are accepted by all given Predicates.

function For_Some (Predicates : Ada_Node_Predicate_Array) → Ada_Node_Predicate : Return a predicate that accepts only nodes that are accepted by at least one of the given Predicates.

function For_All_Children (Predicate : Ada_Node_Predicate; Skip_Null : Standard.Boolean) → Ada_Node_Predicate : Return a predicate that accepts only nodes for which Predicate accepts all children. Unless Skip_Null is false, this does not evaluate the predicate on null children.

function For_Some_Children (Predicate : Ada_Node_Predicate; Skip_Null : Standard.Boolean) → Ada_Node_Predicate : Return a predicate that accepts only nodes for which Predicate accepts at least one child. Unless Skip_Null is false, this does not evaluate the predicate on null children.

function Child_With (Field : Struct_Member_Ref; Predicate : Ada_Node_Predicate) → Ada_Node_Predicate 

Return a predicate that accepts only nodes which have a child corresponding to the given field reference and for which this child is accepted by the given predicate.

Raise a Precondition_Failure if Field is not a valid node field reference.

function Kind_Is (Kind : Ada_Node_Kind_Type) → Ada_Node_Predicate : Return a predicate that accepts only nodes of the given Kind

function Kind_In (First, Last : Ada_Node_Kind_Type) → Ada_Node_Predicate : Return a predicate that accepts only nodes whose kind is in First .. Last.

function Text_Is (Text : Text_Type) → Ada_Node_Predicate : Return a predicate that accepts only nodes that match the given Text

function Node_Is_Null () → Ada_Node_Predicate : Return a predicate that accepts only null nodes

function Decl_Defines (Name : Text_Type) → Ada_Node_Predicate : Return a predicate that accepts only Basic_Decl nodes that define the given Name.

function Xref_Is (Name : Defining_Name; Imprecise_Fallback : Standard.Boolean) → Ada_Node_Predicate : Return a predicate that accepts only nodes whose P_Gnat_Xref property returns Name.

type Ada_Node_Predicate_Array

function Find (Root : Ada_Node'Class; Predicate : access function (N : Ada_Node) return Boolean) → Traverse_Iterator'Class : Return an iterator that yields all nodes under Root (included) that satisfy the Predicate predicate.

function Find (Root : Ada_Node'Class; Predicate : Ada_Node_Predicate'Class) → Traverse_Iterator'Class : Return an iterator that yields all nodes under Root (included) that satisfy the Predicate predicate.

function Find_First (Root : Ada_Node'Class; Predicate : access function (N : Ada_Node) return Boolean) → Ada_Node : Return the first node found under Root (included) that satisfies the given Predicate. Return a null node if there is no such node.

function Find_First (Root : Ada_Node'Class; Predicate : Ada_Node_Predicate'Class) → Ada_Node : Return the first node found under Root (included) that satisfies the given Predicate. Return a null node if there is no such node.

6.5.3. Libadalang.Config_Pragmas

This package provides helpers to deal with GNAT’s configurations pragmas files.

package Unit_Maps is new Ada.Containers.Hashed_Maps

package Unit_Maps is new Ada.Containers.Hashed_Maps
     (Key_Type        => Analysis_Unit,
      Element_Type    => Analysis_Unit,
      Hash            => Hash,
      Equivalent_Keys => "=");

Instantiated generic package:: Unit_Maps

Map analysis unit to the local configuration pragmas file that applies to it.

type Config_Pragmas_Mapping

Components:

Local_Pragmas (Libadalang.Config_Pragmas.Unit_Maps.Map) – Mappings that associate a local configuration pragmas file (element) to each analysis unit (key) for which it applies.
Global_Pragmas (Analysis_Unit) – Configuration pragmas file that applies to all analysis units, or No_Analysis_Unit if there is no such file.

Configuration pragmas file that applies to all analysis units, or No_Analysis_Unit if there is no such file.

procedure Set_Mapping (Context : Analysis_Context; Mapping : Config_Pragmas_Mapping)

Assign in Context configuration pragmas files to analysis units as described in Mapping.

This raises a Precondition_Failure exception if:

Context is null;
any analysis unit in Mapping does not belong to Context;
Local_Pragmas has a No_Analysis key or element.

function Import_From_Project (Context : Analysis_Context; Project : GNATCOLL.Projects.Project_Tree'Class; Subproject : GNATCOLL.Projects.Project_Type) → Config_Pragmas_Mapping 

Load configuration pragmas files referenced in Project and its sub-projects and return a mapping that describes for each analysis unit owned by Project which configuration pragmas files applies to it.

If Subproject is not No_Project, restrict the exploration of local configuration pragmas files to that project (global ones are still found in the root project).

function Import_From_Project (Context : Analysis_Context; Tree : GPR2.Project.Tree.Object; View : GPR2.Project.View.Object) → Config_Pragmas_Mapping 

Load configuration pragmas files referenced in Tree and its sub-projects and return a mapping that describes for each analysis unit owned by Tree which configuration pragmas files applies to it.

If View is not Undefined, restrict the exploration of local configuration pragmas files to that project (global ones are still found in the root project).

procedure Dump (Mapping : Config_Pragmas_Mapping): Dump the content of Mapping on the standard output

procedure Import_From_Project (Context : Analysis_Context; Project : GNATCOLL.Projects.Project_Tree'Class; Subproject : GNATCOLL.Projects.Project_Type): Shortcut for Import_From_Project/Set_Mapping calls

procedure Import_From_Project (Context : Analysis_Context; Tree : GPR2.Project.Tree.Object; View : GPR2.Project.View.Object): Shortcut for Import_From_Project/Set_Mapping calls

6.5.4. Libadalang.Expr_Eval

This package implements an expression evaluator for Libadalang. The aim is to allow at least evaluation of expressions that fit the definitions of static expression in the Ada Reference Manual.

type Big_Integer

type Rational

type Double

type Expr_Kind

type Eval_Result

Discriminants:

Kind (Expr_Kind) –

Components:

Expr_Type (Base_Type_Decl) –
Enum_Result (Enum_Literal_Decl) –
Int_Result (Big_Integer) –
Real_Result (Rational) –
String_Result (Unbounded_Text_Type) – First and Last are used to keep track of the String objects bounds (this information is required to evaluate slices).
First (Standard.Natural) –
Last (Standard.Natural) –

function As_Int (Self : Eval_Result) → Big_Integer : Return the given evaluation result as an Integer, if applicable. This will work for enum or int results, not for real results.

function As_String (Self : Eval_Result) → Unbounded_Text_Type : Return the given evaluation result as a string, if applicable. This will only work for string results.

function Image (Self : Eval_Result) → Standard.String: Return a string representation of Self. Used for testing/debugging purposes.

function Expr_Eval (E : Expr) → Eval_Result : Evaluate the expression passed as parameter

function Expr_Eval_In_Env (E : Expr; Env : Substitution_Array) → Eval_Result : Evaluate the given expression in the context of the given environment

6.5.5. Libadalang.Data_Decomposition

This package provides routines to decompose composite program data into elementary components, and more generally to query the representation of program data.

Attention

This is an experimental feature, so even if it is exposed to allow experiments, it is totally unsupported and the API is very likely to change in the future.

Here is a small example of usage for this package:

--  Code that the program below analyzes, to be compiled with the
--  ``-gnatR4js`` option (``gcc -c pkg.ads -gnatR4js``) to generate the
--  ``pkg.ads.json`` file.

package Pkg is
   type R (N : Natural) is record
      case N is
         when 0 .. 9 =>
            B : Boolean;
         when 100 .. 199 | 900 .. 999 =>
            I : Integer;
         when others =>
            null;
      end case;
   end record;
end Pkg;

--  Corresponding project file

project P is
end P;

--  Program using the ``Libadalang.Data_Decomposition`` API to analyze
--  representation information.

with Ada.Strings.Unbounded; use Ada.Strings.Unbounded;
with Ada.Text_IO;           use Ada.Text_IO;

with GNATCOLL.GMP.Integers; use GNATCOLL.GMP.Integers;
with GPR2.Context;
with GPR2.Path_Name;
with GPR2.Project.Tree;

with Libadalang.Analysis;           use Libadalang.Analysis;
with Libadalang.Common;             use Libadalang.Common;
with Libadalang.Data_Decomposition;
with Libadalang.Iterators;          use Libadalang.Iterators;
with Libadalang.Project_Provider;   use Libadalang.Project_Provider;

procedure Main is
   package DDA renames Libadalang.Data_Decomposition;

   --  Load the "p.gpr" project file, then create the unit provider (to
   --  make Libadalang's name resolution work) and the representation
   --  information collection for it.

   Tree       : GPR2.Project.Tree.Object;
   Provider   : Unit_Provider_Reference;
   Collection : DDA.Repinfo_Collection;
begin
   Tree.Load_Autoconf
     (Filename => GPR2.Path_Name.Create_File
                    (Name      => "p.gpr",
                     Directory => GPR2.Path_Name.No_Resolution),
      Context  => GPR2.Context.Empty);

   Provider := Create_Project_Unit_Provider (Tree);

   Collection := DDA.Load_From_Project (Tree);

   declare
      --  Analyze the "pkg.ads" source file

      Ctx : constant Analysis_Context := Create_Context;
      U   : constant Analysis_Unit := Ctx.Get_From_File ("pkg.ads");

      --  Look for the first type declaration in "pkg.ads"

      function Filter (Node : Ada_Node) return Boolean
      is (Node.Kind in Ada_Base_Type_Decl);

      Decl : constant Base_Type_Decl :=
        Find_First (U.Root, Filter'Access).As_Base_Type_Decl;

      --  Fetch type representation information

      TR : constant DDA.Type_Representation :=
        Collection.Lookup (Decl);

      --  Resolve the layout for this record type for a specific
      --  set of discriminants (N = 950).

      Discs : constant DDA.Discriminant_Values :=
        (1 => Make ("950"));
      RR    : constant DDA.Resolved_Record_Type :=
        DDA.Resolved_Record (TR, Discs);
   begin
      --  Display the size of this record (in bits) as well as the name
      --  and position (in bytes) for each component.

      Put_Line ("Record size:" & RR.Value_Size'Image);
      for C of RR.Components loop
         Put_Line ("Component " & C.Declaration.Image
                   & " at position" & C.Position'Image);
      end loop;
   end;
end Main;

Expected output:

Record size: 64
Component <DefiningName "N" pkg.ads:2:12-2:13> at position 0
Component <DefiningName "I" pkg.ads:7:13-7:14> at position 4

type Numerical_Expression

Expression that, when evaluated, returns an integer (arbitrary precision).

This data type is used to represent dynamic attributes, such as the size of a discriminated record type, which depends on the specific values for its discriminants.

Each numerical expression requires a specific number of integer parameters to be evaluated, each integer corresponding to a record discriminant. See the Discriminant_Count and Evaluate primitives below.

Note that for record types, the Resolved_Record function below resolves all attributes/components at once for a given record type given values for its discriminants. This may be more convenient than going through the evaluation of numerical expressions manually.

No_Numerical_Expression : constant Numerical_Expression

Object type:: Numerical_Expression

function Is_Null (Self : Numerical_Expression) → Standard.Boolean

function Discriminant_Count (Self : Numerical_Expression) → Standard.Natural: Return the number of discriminants needed to evaluate this expression

function Evaluate (Self : Numerical_Expression; Discriminants : Discriminant_Values) → GNATCOLL.GMP.Integers.Big_Integer

Evaluate a numerical expression given specific values for all required discriminants.

Note that more discriminants than required are accepted: the evaluation of the attribute of a component for a record with 2 discriminants may need only the first discriminant that record has

Unsupported_Expression: exception: Exception raised in a function that returns Numerical_Expression when the expression is unsupported. For now, this happens only when the expression depends on dynamic variables that are not discriminants.

type Type_Kind

type Elementary_Type

type Scalar_Type

type Discrete_Type

type Integer_Type

type Real_Type

type Fixed_Type

type Composite_Type

type Type_Representation

Representation information for a given type

No_Type_Representation : constant Type_Representation

Object type:: Type_Representation

function Is_Null (Self : Type_Representation) → Standard.Boolean

function Kind (Self : Type_Representation) → Type_Kind : Return the kind of type Self is

function Alignment (Self : Type_Representation) → Standard.Positive: Alignment for values of this type, in bytes

function Object_Size (Self : Type_Representation) → Numerical_Expression : Number of bits used to hold objects whose type is Self (See GNAT RM 9.6. Value_Size and Object_Size Clauses).

function Value_Size (Self : Type_Representation) → Numerical_Expression : Number of bits required to represent a value whose type is Self. This corresponds to the RM defined 'Size attribute (See GNAT RM 9.6. Value_Size and Object_Size Clauses).

function Scalar_Storage_Order (Self : Type_Representation) → System.Bit_Order: Byte order for scalars stored in this compound type

function Components (Self : Type_Representation) → Component_Representation_Array : List of components for this record, excluding components from the variant part (if any).

function Discriminants (Self : Type_Representation) → Component_Representation_Array : Subset of components for this record that are discriminants

function Has_Variant_Part (Self : Type_Representation) → Standard.Boolean: Whether this record type has a variant part

function Variants (Self : Type_Representation) → Variant_Representation_Array : Assuming this record type has a variant part, return all variants for it

function Bit_Order (Self : Type_Representation) → System.Bit_Order: Order for bit numbering in this record type

function Component_Size (Self : Type_Representation) → Numerical_Expression : Expression that evaluates to the size (in bits) for each array component.

function Small (Self : Type_Representation) → GNATCOLL.GMP.Rational_Numbers.Rational: Return the “small” number for a fixed point type. This is a positive real number, and all values for this fixed point type are multiples of this “small” number.

function Range_First (Self : Type_Representation) → GNATCOLL.GMP.Rational_Numbers.Rational: Lower bound for the mandatory range of a fixed point type

function Range_Last (Self : Type_Representation) → GNATCOLL.GMP.Rational_Numbers.Rational: Upper bound for the mandatory range of a fixed point type

function Resolved_Record (Self : Type_Representation; Discriminants : Discriminant_Values) → Resolved_Record_Type 

Resolve all components in record type Self according to the values of its discriminants.

Raise a Resolution_Error when resolution for the given discriminants yields nonsensical sizes or positions (usually because the discriminants are invalid).

type Component_Representation

Representation information for a given record component

No_Component_Representation : constant Component_Representation

Object type:: Component_Representation

function Is_Null (Self : Component_Representation) → Standard.Boolean

function Declaration (Self : Component_Representation) → Defining_Name : Declaration for this component. This returns No_Defining_Name for artificial components, i.e. components not defined in the source, but introduced by the compiler (for instance the tag for tagged types).

function Component_Name (Self : Component_Representation) → Text_Type : Name for this component. The casing of the result is not specified. This is useful as a complement to the Declaration function above for artificial components, which have no declaration.

function Discriminant_Number (Self : Component_Representation) → Standard.Natural: 0 if this component is not a discriminant. Positive number that uniquely identifies this discriminant in the record type otherwise.

function Position (Self : Component_Representation) → Numerical_Expression : Expression that evaluates to the index of first byte in the record that is used to represent this component.

function First_Bit (Self : Component_Representation) → Standard.Natural: First bit in the byte at Position that is used to represent this component.

function Size (Self : Component_Representation) → Numerical_Expression : Size (in bits) for this component

type Variant_Representation

Representation information for a given record variant (or “variant part alternative”).

No_Variant_Representation : constant Variant_Representation

Object type:: Variant_Representation

function Is_Null (Self : Variant_Representation) → Standard.Boolean

function Present (Self : Variant_Representation) → Numerical_Expression : Expression that evaluates to non-zero when the compoments of the variant are contained in the record type, and to zero when they are not.

function Components (Self : Variant_Representation) → Component_Representation_Array : Components for this variant, excluding components from the sub-variant (if any).

function Has_Subvariant_Part (Self : Variant_Representation) → Standard.Boolean: Whether this variant has a sub-variant part

function Subvariants (Self : Variant_Representation) → Variant_Representation_Array : Assuming this variant has a sub-variant part, return all sub-variants for it.

type Component_Representation_Array

type Variant_Representation_Array

Type_Mismatch_Error: exception: Exception raised when an inconsistency is detected between type representation information and type declarations as found in the sources.

type Discriminant_Values

Actual values for all the discriminants in a record type.

For discriminants that are integers, the value must be the corresponding number. For discriminants that are enumerations, the value must be the code used to represent the enumeration literal (i.e. the equivalent of GNAT’s Enum_Rep attribute). The Discriminant_Value function below can be used to turn a discriminant value into the expected integer value.

No_Discriminant_Value : constant Discriminant_Values

Object type:: Discriminant_Values
Default value:: (1 .. 0 => <>)

Invalid_Discriminant: exception: See the Discriminant_Value function

function Discriminant_Value (Result : Eval_Result) → GNATCOLL.GMP.Integers.Big_Integer: Return the discriminant value corresponding to the given evaluated static expression. Raise an Invalid_Discriminant exception if Result is not an enum literal nor an integer.

type Size_Type

Number of bits or bytes to denote the size of an object in memory.

Note that the range for this type may change in the future, for instance if Ada ever gets ported to a 128-bit architecture.

type Resolved_Component

Components:

Declaration (Defining_Name) – Declaration for this component
Artificial_Name (Unbounded_Text_Type) – If Declaration is null (i.e. if this describes an artificial component), name for that component. The casing is not specified.
Position (Size_Type) – See the corresponding Component_Representation primitive
First_Bit (Standard.Natural) – See the corresponding Component_Representation primitive
Size (Size_Type) – See the corresponding Component_Representation primitive

See the corresponding Component_Representation primitive

type Resolved_Component_Array

type Resolved_Record_Type

Discriminants:

Component_Count (Standard.Natural) –

Components:

Alignment (Standard.Positive) – See the corresponding Type_Representation primitive
Object_Size (Size_Type) – See the corresponding Type_Representation primitive
Value_Size (Size_Type) – See the corresponding Type_Representation primitive
Bit_Order (System.Bit_Order) – See the corresponding Type_Representation primitive
Scalar_Storage_Order (System.Bit_Order) – See the corresponding Type_Representation primitive
Components (Resolved_Component_Array) – See the corresponding Type_Representation primitive

See the corresponding Type_Representation primitive

Resolution_Error: exception: See the Resolved_Record function

type Repinfo_Collection

Collection of compiler-generated representation information

function Lookup (Self : Repinfo_Collection; Decl : Base_Type_Decl'Class) → Type_Representation : Look for the type representation corresponding to the given type declaration. Return No_Type_Representation if nothing in Self matches Type_Name, and raise a Type_Mismatch_Error if an inconsistency is found between Decl and the type representation found for it.

function Load_From_Directory (Name_Pattern : System.Regexp.Regexp; Directory : Standard.String) → Repinfo_Collection : Like Load, but using automatically loading all files whose name matches Name_Pattern in the given Directory.

function Load_From_Project (Tree : GPR2.Project.Tree.Object; View : GPR2.Project.View.Object; Subdirs : Standard.String; Force : Standard.Boolean) → Repinfo_Collection 

Run GPRbuild on the given project Tree to compile all of its Ada units with the -gnatR4js compiler switch, then load all generated JSON files for units under the View sub-project.

Unless left to empty strings, the following formals are passed to gprbuild’s command line:

Subdirs is passed as --subdirs.

If Force is True, pass -f to gprbuild to force the build of compilation units.

Raise a Gprbuild_Error exception if gprbuild exits with a non-zero status code. Raise a Loading_Error exception if the loading of JSON files fails.

Loading_Error: exception: Exception raised when an error occurs while loading a collection of representation information.

type Filename_Array

function Load (Filenames : Filename_Array) → Repinfo_Collection 

Load type representation information from all the given Filenames and return the corresponding collection.

All files are supposed to be generated running GNAT on compilation units with the -gnatR4js switch.

Raise a Loading_Error exception if unsuccessful.

function Load_From_Directories (Name_Pattern : System.Regexp.Regexp; Directories : Filename_Array) → Repinfo_Collection : Like Load, but using automatically loading all files in any of the given Directories whose file name matches Name_Pattern.

Default_JSON_Filename_Regexp : constant GNAT.Regexp.Regexp

Object type:: System.Regexp.Regexp
Default value:: GNAT.Regexp.Compile (".*\.(ad.|a|spc|bdy)\.json")

Default matcher for JSON files created by the -gnatR4js compiler switch. It matches the JSON files created for Ada sources using the most usual file extensions: .ads, .adb, .ada, .spc, .bdy, etc.

Gprbuild_Error: exception: See Load_From_Project

6.5.6. Libadalang.Target_Info

This package provides support for GNAT’s target dependent information files (see GNAT’s -gnatet and gnateT switches).

type Float_Representation

type Digits_Type: Number of digits supported in a floating point type

type Floating_Point_Type_Information

Discriminants:

Present (Standard.Boolean) –

Components:

Digs (Digits_Type) – Number of digits for the floating-point type
Representation (Float_Representation) – Machine representation for the floating-point type
Size (Standard.Positive) – Size in bits for the floating-point type
Alignment (Standard.Natural) – Alignment for the floating-point type

Alignment for the floating-point type

Absent_Floating_Point_Type_Info : constant Floating_Point_Type_Information

Object type:: Floating_Point_Type_Information
Default value:: (Present => False)

function Ada_Type_Definition (Self : Present_Floating_Point_Type_Information) → Text_Type : Return Ada code excerpt that can be used to define a floating point type with the given characteristics.

type Present_Floating_Point_Type_Information

Discriminants:

Present (Standard.Boolean) –

Components:

Digs (Digits_Type) – Number of digits for the floating-point type
Representation (Float_Representation) – Machine representation for the floating-point type
Size (Standard.Positive) – Size in bits for the floating-point type
Alignment (Standard.Natural) – Alignment for the floating-point type

Alignment for the floating-point type

type Floating_Point_Type_Id

List of all the floating-point types that can be described in a Target_Information record.

function C_Name (Self : Floating_Point_Type_Id) → Standard.String

type Mandatory_Floating_Point_Type_Id: Floating-point type that are required in a target information file

type Floating_Point_Types_Information

type Target_Information

Components:

Bits_BE (Standard.Boolean) – Bits stored big-endian?
Bits_Per_Unit (Standard.Positive) – Bits in a storage unit
Bits_Per_Word (Standard.Positive) – Bits in a word
Bytes_BE (Standard.Boolean) – Bytes stored big-endian?
Char_Size (Standard.Positive) – Standard.Character’Size
Double_Float_Alignment (Standard.Natural) – Alignment of double float
Double_Scalar_Alignment (Standard.Natural) – Alignment of double length scalar
Double_Size (Standard.Positive) – Standard.Long_Float’Size
Float_Size (Standard.Positive) – Standard.Float’Size
Float_Words_BE (Standard.Boolean) – Float words stored big-endian?
Int_Size (Standard.Positive) – Standard.Integer’Size
Long_Double_Size (Standard.Positive) – Standard.Long_Long_Float’Size
Long_Long_Long_Size (Standard.Positive) – Standard.Long_Long_Long_Integer’Size
Long_Long_Size (Standard.Positive) – Standard.Long_Long_Integer’Size
Long_Size (Standard.Positive) – Standard.Long_Integer’Size
Maximum_Alignment (Standard.Positive) – Maximum permitted alignment
Max_Unaligned_Field (Standard.Positive) – Maximum size for unaligned bit field
Pointer_Size (Standard.Positive) – System.Address’Size
Short_Enums (Standard.Boolean) – Foreign enums use short size?
Short_Size (Standard.Positive) – Standard.Short_Integer’Size
Strict_Alignment (Standard.Boolean) – Strict alignment?
System_Allocator_Alignment (Standard.Natural) – Alignment for malloc calls
Wchar_T_Size (Standard.Positive) – Interfaces.C.wchar_t’Size
Words_BE (Standard.Boolean) – Words stored big-endian?
Floating_Point_Types (Floating_Point_Types_Information) – Information about floating-point types

Information about floating-point types

Placeholder_Target_Info : constant Target_Information

Object type:: Target_Information
Default value:: (Bits_BE => False, Bits_Per_Unit => 8, Bits_Per_Word => 64, Bytes_BE => False, Char_Size => 8, Double_Float_Alignment => 0, Double_Scalar_Alignment => 0, Double_Size => 64, Float_Size => 32, Float_Words_BE => False, Int_Size => 32, Long_Double_Size => 128, Long_Long_Long_Size => 128, Long_Long_Size => 64, Long_Size => 64, Maximum_Alignment => 16, Max_Unaligned_Field => 1, Pointer_Size => 64, Short_Enums => False, Short_Size => 16, Strict_Alignment => False, System_Allocator_Alignment => 16, Wchar_T_Size => 32, Words_BE => False, Floating_Point_Types => (Float_Id => (Present => True, Digs => 6, Representation => IEEE_754_Binary, Size => 32, Alignment => 32), Double_Id => (Present => True, Digs => 15, Representation => IEEE_754_Binary, Size => 64, Alignment => 64), Long_Double_Id => (Present => True, Digs => 18, Representation => IEEE_754_Binary, Size => 80, Alignment => 128), BF_Id => Absent_Floating_Point_Type_Info, FPRX2_Id => Absent_Floating_Point_Type_Info, HF_Id => Absent_Floating_Point_Type_Info, IF_Id => Absent_Floating_Point_Type_Info, KF_Id => Absent_Floating_Point_Type_Info, RF_Id => Absent_Floating_Point_Type_Info, TF_Id => Absent_Floating_Point_Type_Info, XF_Id => Absent_Floating_Point_Type_Info))

Default target information to use when no specific values were provided. This is just a placeholder: actual values are not guaranteed to remain unchanged across versions.

function Load (Filename : Standard.String) → Target_Information : Read target information from Filename and return it. Raise an Ada.IO_Exceptions.Name_Error or Ada.IO_Exceptions.Use_Error if Filename cannot be read, and raise an Langkit_Support.Errors.Invalid_Input exception if there is any trouble decoding it.

procedure Dump (Self : Target_Information): Debug helper: dump the infomation contained in Self to the standard output.